The irony of this is that these are all very common constraints to graph layout algorithms --- simple type, minimized crossings, label size and placement --- and the label you called out as "better" pessimizes them.
Add to that the fact that Imus' map subtly indicates Kettering's position in a population center (the yellow blob) and gives topographic and foresting information for the area with a clear contrast between the western and eastern sides.
Be careful. The Imus map is deceptively detailed (again: forestation levels), and is not shown in its best light in a PNG clip next to a simplified National Geographic map. You're right; the Imus map looks muddy and harder to read. That's because it's doing more, and is thus not as well suited to a laptop screen (or a clip in a web page) as National Geographic's.
I was impressed with his map.
And then to my surprise I discover that his map is the one on the left!
Maybe in a large printed version his is better? I don't know. All I can tell from the photos is that the computer generated one is better.